Entries listed under “wikipedia”

stephen colbert owns Facebook!

Stephen T Colbert, the host of the Colbert Report and the guy who’s making Man of the Year a true story just rocked Facebook. The social network has this feature where you create a group, and ask other people to join it. The group can be about an agenda, a common peeve, of even a friend’s mother.

Growth is usually viral, people usually learn about a group by reading about it in their newsfeed. The last 10 days have seen one of the fastest growing groups in Facebook history, attracting some press coverage. The 1 million strong for Colbert group is the craziest internet phenomenon I’ve seen in a while.

As the group neared its 1 million mark, I thought it would be a good idea to monitor the growth of the group over the next few hours. With some scripting jiggery-pokery, I was able to record the size of the group every 30 seconds, from 6:30pm yesterday, to 8:30am today morning, which is when I woke up and killed the script.

So here it is, the data, visualized. Presenting to you, the growth of “1,000,000 strong for Colbert”:
chart_facebook

Notice how the growth unsurprisingly curves off as people go to bed at 1am EST (or 11am PST), and begins growing again at 6am.

To provide a complete picture, I also threw in the overall growth data over the days:
overallgrowth

|

startup idea #4984

Here’s an idea I thought of a while ago. You have the storm botnet, which is apparently now capable of being the world’s most powerful supercomputer:

The Storm botnet, or Storm worm botnet, is a massive network of computers linked by the Storm worm Trojan horse in a botnet, a group of “zombie” computers controlled remotely. It is estimated to run on as many as 1,000,000 to 50,000,000 infected and compromised computer systems as of September 2007. Its formation began around January, 2007, when the Storm worm at one point accounted for 8% of all infections on all Microsoft Windows computers.

The botnet reportedly is powerful enough as of September 2007 to force entire countries off of the Internet, and is estimated to be able to potentially execute more instructions per second than some of the world’s top supercomputers.

Obviously, having a large supercomputer is big business these days. So what you do is have a legal version of this. Let’s say you sell computers at 70% of their real price. The only catch is that people will have to run this special software as part of the system. The special software is basically a remote compute client similar to Folding@Home or Google Compute.

Once you have sold enough computers, you essentially have a large army of computers at your beck and call, for 30% the price of what you would have to invest in otherwise. Of course, obviously someone else owns the machines, but while they are doing lightweight tasks such as checking email and chatting, you are folding proteins, running simulations and cracking ciphers.

Now here’s the best part of the deal: the most expensive part of a grid is not the hardware, but the electricity that it uses. And guess who’s paying this electricity! The customer, not you!.

So there you have it. A cheap, one-time cost for an everlasting free CPU grid. Awesome ainnit?

note: This idea is under this license.

The Big House

I’ve been going to the home games this season, thought I’d post pictures. It should be noted that The Big House (the nickname for the Michigan Stadium) is the largest American football stadium and the fifth largest stadium in the world.

tech interview preparation for non-cs people

A non-computer science person emailed me that she had an interview with a dotcom company next week, and wanted to know what she could do to prepare for it. I ended up writing a fairly sizable email, which I thought would be useful to others as well. Here’s my response:

I am assuming they have an idea about your background. If they know that you are not a CS person, then they should try to ask you questions that test your problem-solving abilities overall. Additionally, I guess they will ask you some basic CS questions so that they know that you can pick up CS skills when you’re hired.

Some questions are also posed as logic questions, but actually have a deep CS background. For example: “when you are doing laundry, and you need to fold your socks, and you have a big tabletop to put stuff on, what’s the best way to arrange them in matching pairs?”. One option is to take a sock, and then take another one. If they match, then it’s good. Else, throw the socks back in, and start again. Obviously, this is a bad solution. Look for “bucket sort” for a good solution.

You can compensate for this by just being very good at logic, but it does help to know some CS. Try to have an idea about computational complexity. You can ask any CS undergrad (senior, etc) about this.

1. http://pages.cs.wisc.edu/~hasti/cs367-common/notes/COMPLEXITY.html
2. http://en.wikipedia.org/wiki/Computational_complexity_theory

It helps to run through a few questions so that your brain is ready for the questions during the interview.

For problem solving, check out
http://www.techinterview.org/

I googled “dotcomcompanyname interview questions” and found some more links.

Hope that helps, Good luck!
-a